INDEX

Explanations

phrases indicating a call to action or instruction

oai_token-act-pair · gpt-3.5-turbo

phrases that begin with "let's."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 holiest

-0.77

 Languages

-0.68

 Soldier

-0.67

 cumbers

-0.67

oppable

-0.66

CLASSIFIED

-0.65

 millenn

-0.64

manship

-0.64

atana

-0.63

AGES

-0.58

POSITIVE LOGITS

icia

0.89

tered

0.88

tering

0.80

itia

0.80

hetically

0.75

us

0.74

 slip

0.74

uls

0.72

ersen

0.69

enne

0.68

Activations Density 0.029%

phrases indicating a call to action or instruction

phrases that begin with "let's."

No Comments

No Known Activations